a photo    

赵 瑞辉 (Ruihui, Zhao)

腾讯天衍实验室 高级研究员(T10)
Senior Researcher
Tecent Jarvis Lab

Email: zacharyzhao [AT] tencent [DOT] com
Address: Tencent Building, 10000 Shennan Boulevard Nanshan District, Shenzhen 518000, China

Biography

Ruihui Zhao received his B.S. degree at UESTC (2015) and M.S. degree at Waseda University (早稻田大学, 2017). He joined Tencent Jarvis Lab as a senior researcher in early 2018. Previously, he was an NLP engineer at Sinovation Ventures (2017 - 2018). He has several papers accepted by ACL, WWW, AAAI, NAACL, IJCNN, TOIT, IEEE WCSP, etc. His research and projects mainly focus on NLP and information security.


We are looking for FTE & Research interns who are interested in Healthcare AI. Please send your CV to me if you want to join us.

Research Interests

  • Natural Language Processing: Medical Pre-training Language Modeling, Knowledge Distillation, Text Matching, Taxonomy Construction and Expansion, Question Generation and Answering, etc.
  • Machine Learning Techniques: Time Series Anomaly Detection, Community Detection, Graph Embeddings, Model Interpretability, etc.
  • Information Security: Federated Learning, Searchable Encryption
  • Co-advised Research Interns

  • Wang Suyuchen (Ph.D. student, Univ. of Montreal, 2020.07-2021.07)
  • Cheng Yi (Bachelor, Peking Univ., 2020.06-)
  • Wu Sifan (Bachelor, Tsinghua Univ., 2021.4-)
  • Ou Zijing (Bachelor, Sun Yat-sen Univ., 2020.05-2021.06)
  • Liu Yi (Ph.D. student, Monash Univ., 2020.06-2021.01)
  • Hou Zhengxu (Master, USC, 2020.05-2021.2)
  • Li Siyao (Master student, CMU, 2020.06-2020.08)

  • Projects

    Applying NLP and machine learning techniques to the fields of healthcare and security:
  • Lead a healthcare assistant project and more than 6 full-time research engineers;
  • Lead a medical insurance anti-fraud project and 5 full-time research engineers, including techniques such as gang detection, time series anomaly detection, graph embeddings, model interpretability, federated learning, etc.
  • Owned a medical QA and search ranking system which supported Tencent Healthcare;
  • Owned a federated machine learning project, cooperated with Webank AI team;
  • Owned a medical entity recommendation project based on NER, intent detection and medical ontology, cooperated with Tencent Wechat search team;
  • Owned a project which applied explainable anomaly detection algorithms on sequential security data, cooperated with Tencent Security team;
  • Owned a Language Model toolkit, cooperated with Tencent PCG Venus platform;
  • Participated in an intelligent aided diagnosis system.

  • Selected Publication [Google Scholar]

    2021

    Refining BERT Embeddings for Document Hashing via Mutual Information Maximization
        Zijing Ou, Qinliang Su, Jianxing Yu, Ruihui Zhao, Yefeng Zheng and Bang Liu
        The 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP-2021 findings, CCF-B), Punta Cana, Dominican Republic

    Integrating Semantics and Neighborhood Information with Graph-Driven Generative Models for Document Retrieval
        Zijing Ou, Qinliang Su, Jianxing Yu, Bang Liu, Jingwen Wang, Ruihui Zhao, Changyou Chen and Yefeng Zheng
        The 59th Annual Meeting of the Association for Computational Linguistics (ACL-2021, CCF-A), Bangkok, Thailand

    Guiding the Growth: Difficulty-Controllable Question Generation through Step-by-Step Rewriting
        Yi Cheng, Siyao Li, Bang Liu, Ruihui Zhao, Sujian Li, Chenghua Lin and Yefeng Zheng
        The 59th Annual Meeting of the Association for Computational Linguistics (ACL-2021, CCF-A), Bangkok, Thailand

    Imperfect also Deserves Reward: Multi-Level and Sequential Reward Modeling for Better Dialog Management
        Zhengxu Hou, Bang Liu, Ruihui Zhao, Zijing Ou, Yafei Liu, Xi Chen, Yefeng Zheng
        2021 Annual Conference of the North American Chapter of the Association for Computational Linguistics (NAACL-2021, NLP top-tier Conf.), Virtual Conference

    Enquire One's Parent and Child Before Decision: Fully Exploit Hierarchical Structure for Self-Supervised Taxonomy Expansion
        Suyuchen Wang, Ruihui Zhao, Xi Chen, Yefeng Zheng, Bang Liu
        The Web Conference (WWW-2021, CCF-A), Virtual Conference

    FedOVA: One-vs-All Training Method for Federated Learning with Non-IID Data
        Yuanshao Zhu, Christos Markos, Ruihui Zhao, Yefeng Zheng, James Yu
        International Joint Conference on Neural Networks (IJCNN-2021, CCF-C), Virtual Conference

    Graph-Evolving Meta-Learning for Low-Resource Medical Dialogue Generation
        Shuai Lin, Pan Zhou, Xiaodan Liang, Jianheng Tang, Ruihui Zhao, Ziliang Chen, Liang Lin
        The Thirty-Fifth AAAI Conference on Artificial Intelligence (AAAI-2021, CCF-A), Virtual Conference
        [Codes]

    Towards Communication-efficient and Attack-resistant Federated Edge Learning for Industrial Internet of Things
        Yi Liu, Ruihui Zhao, Jiawen Kang, Abdulsalam Yassine, Dusit Niyato, Jialiang Peng*
        Transactions on Internet Technology (TOIT-2021, CCF-B, IF=4.79)

    2020

    Privacy-Preserving Technology to Help Millions of People: Federated Prediction Model for Stroke Prevention
        Ce Ju*, Ruihui Zhao*, Jichao Sun, Xiguang Wei, Bo Zhao, Yang Liu, Hongshan Li, Tianjian Chen, Xinwei Zhang, Dashan Gao, Ben Tan, Han Yu and Yuan Jin
        The 29th International Joint Conference on Artificial Intelligence (FML'IJCAI-2020), Virtual Conference

    Forecasting the long-term trend of COVID-19 epidemic using a dynamic model
        Jichao Sun,† Xi Chen,† Ziheng Zhang,† Shengzhang Lai, Bo Zhao, Hualuo Liu, Shuojia Wang, Wenjing Huan, Ruihui Zhao, Man Tat Alexander Ng*, Yefeng Zheng*
        Scientific Reports (Scientific Reports, IF=4.525)

    2019 and Before

    A Lightweight Efficient Searchable Encryption Scheme using Supervised Sentence Representations
        Ruihui Zhao, Mizuho Iwaihara, etc.
        The 28th International Joint Conference on Artificial Intelligence (FL'IJCAI-2019), Macau, China

    A Community-based P2P OSNs Using Broadcast Encryption Supporting Cross-platform with High-Security
        Mingjie Ding, Ruihui Zhao, Keiichi Koyanagi, Takeshi Tsuchiya, Hiroaki Sawano
        8th International Conference on Wireless Communications & Signal Processing (IEEE WCSP-2016, EI), Yangzhou, China

    Privacy-preserving personalized search over encrypted cloud data supporting multi-keyword ranking
        Ruihui Zhao, Hongwei Li, Yi Yang, Yu Liang
        6th International Conference on Wireless Communications & Signal Processing (IEEE WCSP-2014, EI), Hefei, China


    Academic Services

  • Industry Chair: AAAI 2021 rseml workshop, [Link]
  • Program Committee Board (PCB) member of IJCAI (serve for IJCAI-ECAI 2022, IJCAI 2023 and IJCAI 2024)
  • PC member: IJCAI 2021
  • Reviewer: EMNLP 2021, TPDS 2018 (CCF-A Trans.), AACL 2020, etc.

  • Granted Patents

    9 Chinese patents granted, another ~70 patents pending
  • 2021.07 一种基于口语化同义句生成的数据增强系统, 赵瑞辉/乔倩倩/韦伟, CN 110110045 A
  • 2021.06 一种适用于问答系统的细粒度树形分类器构建方案, 赵瑞辉/韦伟/乔倩倩/谭雯雯, CN 110032631 A
  • 2021.03 基于神经网络语言模型的轻量级加密云数据检索方案, 赵瑞辉/乔倩倩/许顺楠, CN 109992978 A
  • 2021.02 一种基于因果图以及深度学习的疾病预测诊断方法, 赵瑞辉/王婧雯, CN 112035671 A
  • 2021.01 一种通信高效的基于异步梯度压缩的联邦学习框架, 刘毅/赵瑞辉, CN 111784002 A
  • 2020.11 一种基于多头自注意力的知识蒸馏的语言模型压缩方法, 黄展鹏/赵博/赵瑞辉/陆扩建, CN 111554268 A
  • 2020.10 一种基于预构建本体和知识图谱的可控医疗实体推荐方案, 赵瑞辉/陆扩建/赵博/黄展鹏, CN 111538894 A
  • 2020.10 一种可解释的基于知识蒸馏的轻量级搜索排序方案, 赵瑞辉/陆扩建/赵博/黄展鹏, CN 111538908 A
  • 2020.06 一种综合纵向-横向联邦学习的分散数据协同训练方案, 李洪珊/赵瑞辉/赵博, CN 111081337 A

  • News, Honors and Awards

  • 2021.11 牵头天衍实验室与同济大学助理教授程大伟在医保基金反欺诈领域的高校合作
  • 2021.10 获得中国计算语言学大会(CCL 2021)评测任务3:智能医疗对话诊疗评测智能化医疗诊断赛道一等奖 [Link]
  • 2021.10 与中山大学副教授梁小丹合作的课题3.3医疗自然语言理解获得2020年CCF-腾讯犀牛鸟基金优秀奖(6/25)
  • 2021.10 主持2021 CCF-腾讯犀牛鸟基金课题4.3医疗自然语言理解, 牵头与上海交通大学助理教授陈露的科研合作
  • 2021.07 成功联合主办ICLR MLPCP challenge, 助推医疗对话生成和辅助诊断的行业交流与发展, [Link1], [Link2], [Link3]
  • 2021.05 数博会-领先科技成果奖(撰写人&第二完成人, 国家社会科技奖, 获奖比例: 49/560,路演比例: 5/49), [Link1], [Link2]
  • 2021.05 ACL录用新闻报道 by 机器之心, [Link]
  • 2021.05 联合主办并赞助CCKS评测, [Link]
  • 2021.03 荣获腾讯“年度专利接口人”(top-10, 腾讯云与智慧产业事业群唯一, 带领天衍实验室在2020年度共申请了146项发明专利,增长53.7%)
  • 2021.01 作为项目Sponsor及技术负责人,研发国内首个新冠疫苗智能问答(已上线到腾讯健康及微信搜一搜), [Link]
  • 2020.12 荣获腾讯“业务突破奖”、腾讯“同心战役特别贡献奖”
  • 2020.11 南山区“十大创新工匠”提名奖(top-20,与南山区100多家企业如大疆、中兴、大族激光、中国建筑集团等公司的创始人、CEO、首席科学家等同台竞技), [Link1], [Link2]
  • 2020.10 主持2020 CCF-腾讯犀牛鸟基金课题3.3医疗自然语言理解, 牵头与中山大学副教授梁小丹的科研合作, [Link]
  • 2020.09 首届北京数智医保创新竞赛金奖 (技术负责人), [Link]
  • 2020.08 作为Sponsor,和杨强刘洋等牵头成立腾讯天衍-微众联邦学习联合实验室, [Link]
  • 2020.04 与微众银行AI部门联合研发并发布业界首个医疗联邦学习应用框架, [Link]
  • 2020.03 牵头天衍实验室与加拿大蒙特利尔大学助理教授刘邦的科研顾问合作
  • 2020.03 作为腾讯代表参与IEEE Standard P3652.1-Federated Machine Learning首个联邦学习国际标准制定会议, [Link]
  • 2020.03 "Explainable Anomaly Detection Using Spark", Ruihui Zhao, Ting Chen, presentation accepted by Spark AI Summit 2020
  • 2020.02 担任腾讯医疗与钟南山院士团队成立的大数据及人工智能联合实验室核心研究员,并进行新冠重症预测及流感预测方面的研究